Start-Ups Work to Curb
Junk e-mail represents the dark side of the Internet. But there could be a silver lining -- the chance for some fighters of such "spam" to get rich.
As other Web-based business ideas fade, the prospect of selling software and services to block unwanted e-mail has become one of the hottest concepts attracting entrepreneurs and venture capitalists. In the San Francisco area alone, at least four young companies have popped up to promote technology that lets users control what messages reach them.
"It's a no-brainer to me as an investment," said Joe DiDonato, an "angel" investor who put money into start-up Cloudmark Inc. He says spam "has to be one of the largest problems in the world. Anytime you have something that large ... it's a killer investment."
The start-ups are using a variety of approaches to curtail spam, some representing sharp departures from the past. For example, Cloudmark, which plans to release its software to consumers Wednesday, is exploiting some of the same principles used in Napster Inc.'s music-sharing service.
Many experts believe that spam can never be eliminated, in part because spammers constantly devise new ways to send their annoying solicitations around the Internet. Some of them, for example, use sophisticated techniques to disguise their identities and the origination point of their messages, two attributes that are frequently used in filtering out spam.
"There's no perfect way to solve this problem," said Laura Koetzle, an analyst at Forrester Research Inc. in Cambridge, Mass. "You have to deal with the downsides of any method you choose."
Up to now, antispam software has generally sought to single out unwanted messages. This technique is sometimes called a "blacklist," because it uses a database of known spam. By comparing incoming messages against this list of suspect mail, software can filter out the bad e-mail from the good.
Brightmail Inc., a closely held San Francisco company, has created a twist on the blacklist concept. While blacklists often use a set of static rules about how to identify spam, Brightmail's technology constantly updates those rules based on new spam, and then pushes them out to the computers of customers running its software. There the spam is filtered and blocked when it hits the user's mail gateway.
The company also boasts a full-time data center of people monitoring the company's system to make adjustments as spam attacks pop up. The start-up, which boasts among its clients Internet services such as EarthLink Inc. and Microsoft Corp.'s MSN service, recently attracted a new chief executive, Enrique Salem, a well-regarded software-industry veteran.
A newer concept is to identify and accept only e-mail from legitimate senders, automatically filtering everything else out. This "whitelist" approach is being used by start-ups MailShell .com Inc. of Santa Clara, Calif., and IronPort Systems Inc. of San Bruno, Calif., though their approaches differ.
MailShell's software assigns a "spam score" for each piece of e-mail based on certain criteria, such as if it is in a database of known spammers or matches certain key words. If a sender is included on MailShell's list of approved senders, such as the user's United Airlines frequent-flier program, then the e-mail's "spam score" decreases. The e-mail will reach the user's mailbox depending on parameters that the user sets and the e-mail's score.
IronPort sets up a market mechanism that asks marketers to put up a financial "bond" to be placed on its white list and to guarantee to IronPort that their e-mail isn't spam. IronPort will deduct a "fine" from a company's bond if it is found to send out spam. IronPort is betting that spammers won't be willing to put up -- and potentially lose -- such money.
Cloudmark uses Napster-like "peer-to-peer" technology, based on getting many individual computers on the Internet to work together. The San Francisco company assigns a kind of signature to pieces of spam, and forwards its contents to other far-flung computers on the Internet that are running its software. The signatures help block duplicate copies of the same message from spreading.
The company relies on individual users to "vote" on whether an e-mail is or isn't spam; a "yes" vote assigns a signature to the spam and forwards it to a computer network to be tallied. Cloudmark, whose consumer software will be free, recently recruited Karl Jacob, a former Microsoft executive who had recently run Keen Inc.
Yet another private company, Postini Corp. of Redwood City, Calif., offers a service that routes companies' or Internet services' e-mail through its computer system to filter out spam, using a set of statistical techniques to define it. It returns only the remaining screened e-mail back to the company or Internet service provider.
These new approaches all have drawbacks. Most are designed for corporations and Internet service providers; consumers who aren't employees or users of those services won't get the benefits.
Brightmail, which has proved to be quite effective according to customers like EarthLink, "works acceptably but not miraculously well," said Forrester's Ms. Koetzle. Brightmail's performance may bog down as its software scans through a larger and larger stack of rules to identify and separate spam, making it harder to keep up with exponential increase in spam volumes, the analyst warns.
White-list approaches like MailShell and IronPort "can be effective in a limited sense," said Michael Osterman, founder of Osterman Research, an independent research firm in Black Diamond, Wash., that specializes in messaging and directory technology. But some corporations and Internet-access services don't want to close themselves off to new customers or other contacts. The list of accepted senders must be updated quickly enough for users to receive their mail and the likelihood of "false positives," or e-mail which resembles spam but actually contains critical correspondence, may be high, Mr. Osterman said.
Cloudmark cites its community of users who vote on spam as one of its biggest strengths, since it receives their services free. But Forrester's Ms. Koetzle pointed out that may be its biggest weakness since users may have little to motivate them to participate.
These technologies are likely to reduce rather than eliminate spam, in part because spam costs so little to send and there are so many people with an interest in finding ways to evade technological limits. But just as immunizations can help biological infections, better technology might make a noticeable difference. "It might not actually stop but it'll get better," Mr. Osterman said.
Write to Mylene Mangalindan at email@example.com
Updated June 19, 2002 4:28 p.m. EDT
To distribute multiple copies of this article, visit the Dow Jones Reprints site.
Corrections Contact Us Help About Dow Jones Mobile Devices
Copyright © 2002 Dow Jones & Company, Inc. All Rights Reserved
Copyright and reprint information